In various situations, it is useful to convert text to machine-encoded text. Oftentimes, to perform optical character recognition, a page containing printed text is electronically captured using a flatbed scanner, mounted camera, or some other equivalent device. The electronically captured image of the printed page is processed to create machine-encoded text that is an accurate representation of the information contained in the printed text. However, in some circumstances, the printed text is not scanned using a fixed position device. Rather, a handheld device, such as a camera, is used to capture an image of the text that is to be converted to machine-encoded text. The use of a handheld device to capture images can result in blurred pictures. This blurring is often due to unintentional movement of a user's hand, arm, and/or body while the user is attempting to hold the handheld device still at the time an image is being captured. As the amount of time to capture an image increases (e.g., the longer the camera's shutter is open), the amount of blurring in the image can be expected to increase. While some techniques for reducing blur in captured images exist, these techniques tend to require significant computing power and can be difficult to execute efficiently on a handheld device due to factors such as the device's processing capabilities and its power source.